A method for positioning a problem of a defective product in an electrolytic cell assembly process

Abstract

The application relates to a problem positioning method of defective products in an electrolytic cell assembly process, aiming at solving the technical problem of low problem positioning efficiency and positioning accuracy of existing defective products. The application provides a problem positioning method of defective products in an electrolytic cell assembly process, which determines whether a product to be detected is a defective product by constructing a standard parameter library and combining corresponding judgment rules, and positions problems of the defective product by constructing an association rule library and combining corresponding judgment rules.

Description

Technical Field

[0001] This invention relates to the problem location of defective products, and more specifically to a method for locating the problem of defective products in the electrolytic cell assembly process. Background Technology

[0002] In the electrolytic cell assembly process, the identification of defective products plays a crucial role in bridging the gap between upstream and downstream processes. By accurately identifying and defining the triggering points of defective products in the assembly process, the recurrence of similar problems can be avoided from the source, thereby reducing the quality risks associated with each process. Simultaneously, the identification of defective products provides a clear basis for quality judgment throughout the entire electrolytic cell assembly process, effectively improving the stability and controllability of the assembly process. Therefore, the identification of defective products becomes a core link in achieving closed-loop quality control in electrolytic cell assembly.

[0003] Traditional methods for locating problems with defective products often rely on the experience and judgment of technicians. Technicians disassemble and analyze each parameter of the defective product one by one. This process is not only time-consuming and labor-intensive, but the accuracy of the location results also depends on the professional level and experience of the technicians. Misjudgments or omissions are prone to occur, resulting in low efficiency and accuracy in locating problems with defective products. Summary of the Invention

[0004] The purpose of this invention is to solve the technical problems of low efficiency and accuracy in locating defective products in existing methods, and to provide a method for locating defective products in the electrolytic cell assembly process.

[0005] To achieve the above objectives, the technical solution provided by this invention is as follows:

[0006] A method for locating defective products in the assembly process of an electrolytic cell, characterized by the following steps:

[0007] S1. Construct an association rule base, which is used to store historical defective product data. The historical defective product data includes defective product types, inspection data of each type of defective product, and the types of problems existing in each type of defective product.

[0008] S2. Obtain the test parameter values ​​of the product to be tested in the electrolytic cell assembly process. Based on the pre-built standard parameter library, determine whether the deviation values ​​between each test parameter value and the corresponding design value are all within the design tolerance range. If so, the product to be tested is a good product, and each test parameter is marked as a normal parameter and recorded. If not, the product to be tested is a defective product. Extract the abnormal parameters from the test parameters and record them, then execute step S3. The design values ​​and design tolerance ranges are stored in the standard parameter library.

[0009] S3. Compare the defective products and their abnormal parameters with the defective product types and inspection data in the association rule base, and determine the candidate problem types of the defective products based on the comparison results and problem types; according to the comparison results, sort the candidate problem types from high to low confidence level, and take the candidate problem type with the highest confidence level as the current candidate problem type of the defective product.

[0010] S4. After adjusting the corresponding assembly process according to the current candidate problem type, reprocess the defective products and obtain the new detection parameter values ​​of the defective products.

[0011] S5. Determine whether the deviation values ​​between each new detection parameter value and the corresponding design value are all within the design tolerance range. If not, determine that the candidate problem type is not a problem of the defective product, take the next level candidate problem type of the current candidate problem type as the current candidate problem type of the defective product, and return to step S4. If yes, determine that the current candidate problem type is a problem of the defective product, thereby completing the problem location of the defective product.

[0012] Furthermore, step S2 specifically includes:

[0013] S2.1 Construct a standard parameter library, which stores the design parameters of each product in the electrolytic cell assembly process, their corresponding design values ​​and design tolerance ranges.

[0014] S2.2 Obtain the values ​​of various test parameters for the product to be tested during the electrolytic cell assembly process;

[0015] S2.3. Based on the type of product to be tested, determine the corresponding design parameters from the standard parameter library, and then determine the corresponding design values ​​and design tolerance ranges. Compare the test parameter values ​​with the design values ​​to obtain the deviation values. Determine whether each deviation value is within the design tolerance range. If so, the product to be tested is a good product. Mark each test parameter as a normal parameter and record the normal parameter values. If not, the product to be tested is a defective product. Extract the abnormal parameters from the test parameters and record them, then proceed to step S3.

[0016] Furthermore, in step S1, the construction process of the association rule base is as follows:

[0017] S1.1, Construct the initial association rule base;

[0018] S1.2 Collect historical data on defective products and construct a training dataset;

[0019] S1.3. Using the training dataset as input, the decision tree algorithm is used to train the initial association rule base and output the trained association rule base, thus completing the construction of the association rule base.

[0020] Furthermore, in step S5, if none of the candidate problem types in the association rule base are problems related to defective products, then after manually identifying the new problem of the defective product, the new problem is recorded. When the new problem and the identified problems of defective products accumulate to a preset number, they are updated to the training dataset, and the association rule base is retrained to complete the update of the association rule base.

[0021] Furthermore, in step S2.2, the values ​​of various detection parameters of the product to be tested in the electrolytic cell assembly process are obtained by using docking sensors or manually.

[0022] Furthermore, in step S2.3, when the product to be inspected is defective, if there is a critical inspection parameter whose deviation value is exactly the critical value, then the abnormal parameter and the critical inspection parameter are extracted and recorded, and then step S3 is executed. During the execution process, the critical inspection parameter is used as a reference to determine the candidate problem type; the critical value is one of the values ​​at both ends of the design tolerance range.

[0023] The beneficial effects of this invention are:

[0024] 1. This invention provides a method for locating defects in the assembly process of an electrolytic cell. By constructing a standard parameter library and combining it with corresponding judgment rules, it determines whether the product to be inspected is defective. By constructing an association rule library and combining it with corresponding judgment rules, it locates the problem of the defective product. The above-mentioned automated process replaces the traditional manual analysis, reduces human intervention, shortens the time for locating defects, and avoids the problem of low efficiency and accuracy in locating defects due to differences in the ability and experience of quality analysts.

[0025] 2. This invention provides a method for locating defective products in the assembly process of an electrolytic cell. By accumulating historical data on defective products and updating the association rule base, it identifies potential risks in the production process, transforming quality problems from "post-event handling" to "pre-event prevention," thereby improving product quality.

[0026] 3. The present invention provides a method for locating defective products in the assembly process of an electrolytic cell, which uses an incremental learning-based association rule base update method to ensure that the association rule base is adapted to different product types and changes in production processes. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating an embodiment of a method for locating defective products in an electrolytic cell assembly process according to the present invention.

[0028] Figure 2 This is a schematic diagram of the comparative analysis process of various detection parameters in step S2.3 of an embodiment of the method for locating defective products in the assembly process of an electrolytic cell according to the present invention. Detailed Implementation

[0029] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0030] This embodiment describes a method for locating defective products in the electrolytic cell assembly process, using quality inspection data from the electrolytic cell assembly process as an example. Figure 1 As shown, it includes the following steps:

[0031] S1. Construct an association rule base, which is used to store historical defective product data. The historical defective product data includes defective product types, inspection data of each type of defective product, and the types of problems existing in each type of defective product.

[0032] The construction process of the association rule base is as follows:

[0033] S1.1, Construct the initial association rule base;

[0034] S1.2 Collect historical data of defective products from the past 3 years, filter the historical data, delete historical data with no problem type, missing inspection parameters or an anomaly rate greater than 5%, and then construct a training dataset.

[0035] S1.3. Using the training dataset as input, the decision tree algorithm is used to train the initial association rule base. Training stops when the accuracy is greater than 95%, and the trained association rule base is output.

[0036] S2. Obtain the test parameter values ​​of the product to be tested in the electrolytic cell assembly process. Based on the pre-built standard parameter library, determine whether the deviation values ​​between each test parameter value and the corresponding design value are all within the design tolerance range. If so, the product to be tested is a good product. Mark each test parameter as a normal parameter and record it. If not, the product to be tested is a defective product. Extract the abnormal parameters from the test parameters and record them. Then execute step S3. The design values ​​and design tolerance ranges are stored in the standard parameter library.

[0037] S2.1 Construct a standard parameter library, which stores the design parameters of each product in the electrolytic cell assembly process, their corresponding design values ​​and design tolerance ranges.

[0038] For quality inspection data in the electrolytic cell assembly process, multiple inspection procedures are set up, each corresponding to multiple inspection items. For example, the stamping process includes inspections of appearance, dimensions, and symmetry; the welding process includes weld height, undercut length, depth, and surface defects. The procedures and parameters are numbered. For instance, the stamping process is set as G01, the welding process as G02, the appearance parameter as C001, the length parameter as C002, the width parameter as C003, and the weld height as C004. Design values ​​and tolerance ranges for each component's parameters (diameter, length, height, weld quality, etc.) are set, and the parameters for each component are prioritized: Level 1 represents important parameters, Level 2 represents minor parameters, and Level 3 represents general parameters. Appearance inspection parameters are judged according to severity levels, divided into four levels: Level 1 is severe, and Level 4 has no impact. The resulting records are shown in Table 1.

[0039] Table 1

[0040]

[0041] S2.2 Obtain the values ​​of various test parameters for the product to be tested during the electrolytic cell assembly process; when the process reaches the current test step, the inspector selects the corresponding step number and parameter and records the values ​​manually or directly by data acquisition equipment (if equipped with a data transmission measuring device). The resulting data is shown in Table 2.

[0042] Table 2

[0043]

[0044] S2.3, such as Figure 2 As shown, based on the type of product to be inspected, the corresponding design parameters are determined from the standard parameter library, and then the corresponding design values ​​and design tolerance ranges are determined. The inspection parameter values ​​V1 and design values ​​V2 are compared to obtain the deviation values ​​Δ1. It is determined whether each deviation value is within the design tolerance range. If so, the product to be inspected is a good product, and each inspection parameter is marked as a normal parameter and its value is recorded. If not, the product to be inspected is a defective product. Abnormal parameters are extracted from the inspection parameters and recorded, and then step S3 is executed. When the product to be inspected is a defective product, if there is a critical inspection parameter whose corresponding deviation value is exactly the critical value, then the abnormal parameter and the critical inspection parameter are extracted and recorded simultaneously, and then step S3 is executed. During the execution process, the critical inspection parameter is used as a reference for determining the candidate problem type. The critical value is one of the values ​​at either end of the design tolerance range. Parameters are divided into appearance parameters and numerical parameters. When a parameter is an appearance parameter, its appearance inspection defects, etc., have no numerical value and are judged according to the defect level. The resulting comparison records are shown in Table 3.

[0045] Table 3

[0046]

[0047] S3. Compare the defective products and their abnormal parameters with the defective product types and inspection data in the association rule base. Based on the comparison results and problem types, determine the candidate problem types for the defective products. According to the comparison results, sort the candidate problem types from high to low confidence level, and take the candidate problem type with the highest confidence level as the current candidate problem type for the defective product. That is, use the data collected from the defective products to predict according to the trained association rule base to obtain a list of candidate problem types and the confidence level corresponding to each candidate problem type.

[0048] S4. After adjusting the corresponding assembly process according to the current candidate problem type, reprocess the defective products and obtain the new detection parameter values ​​of the defective products.

[0049] S5. Determine whether the deviation between each new detection parameter value and the corresponding design value is within the design tolerance range. If not, determine that the candidate problem type is not a problem of the defective product. Take the next level candidate problem type of the current candidate problem type as the current candidate problem type of the defective product and return to step S4. If yes, determine that the current candidate problem type is a problem of the defective product, thereby completing the problem location of the defective product. At the same time, record the modification process and problem of the defective product.

[0050] If the problem type is "Welding speed too slow," verify the current welding speed and the weld data after adjusting the welding speed. If the weld height is acceptable after adjusting the welding speed, then the problem of the defective product is determined to be "Welding speed too slow." Otherwise, proceed to verify the next candidate problem. If none of the candidate problem types are valid, manual judgment can be performed, and the final problem result can be entered.

[0051] If none of the candidate problem types in the association rule base are related to defective products, then a new problem for the defective product is manually identified and recorded. When the number of new problems and identified defective product problems accumulates to 100, the data is updated to the training dataset, and the association rule base is retrained to complete the update of the association rule base.

[0052] Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this application.

Claims

1. A method for locating defective products in the assembly process of an electrolytic cell, characterized in that, Includes the following steps: S1. Construct an association rule base, which is used to store historical defective product data. The historical defective product data includes defective product types, inspection data of each type of defective product, and the types of problems existing in each type of defective product. S2. Obtain the test parameter values ​​of the product to be tested in the electrolytic cell assembly process. Based on the pre-built standard parameter library, determine whether the deviation values ​​between each test parameter value and the corresponding design value are all within the design tolerance range. If so, the product to be tested is a good product, and each test parameter is marked as a normal parameter and recorded. If not, the product to be tested is a defective product. Extract the abnormal parameters from the test parameters and record them, then execute step S3. The design values ​​and design tolerance ranges are stored in the standard parameter library. S3. Compare the defective products and their abnormal parameters with the defective product types and inspection data in the association rule base, and determine the candidate problem types of the defective products based on the comparison results and problem types. Based on the comparison results, the candidate problem types are sorted from high to low confidence level, and the candidate problem type with the highest confidence level is taken as the current candidate problem type of the defective product. S4. After adjusting the corresponding assembly process according to the current candidate problem type, reprocess the defective products and obtain the new detection parameter values ​​of the defective products. S5. Determine whether the deviation values ​​between each new detection parameter value and the corresponding design value are all within the design tolerance range. If not, determine that the candidate problem type is not a problem of the defective product, take the next level candidate problem type of the current candidate problem type as the current candidate problem type of the defective product, and return to step S4. If yes, determine that the current candidate problem type is a problem of the defective product, thereby completing the problem location of the defective product.

2. The method for locating defective products in the electrolytic cell assembly process according to claim 1, characterized in that, Step S2 is as follows: S2.1 Construct a standard parameter library, which stores the design parameters of each product in the electrolytic cell assembly process, their corresponding design values ​​and design tolerance ranges. S2.2 Obtain the values ​​of various test parameters for the product to be tested during the electrolytic cell assembly process; S2.

3. Based on the type of product to be tested, determine the corresponding design parameters from the standard parameter library, and then determine the corresponding design values ​​and design tolerance ranges. Compare the test parameter values ​​with the design values ​​to obtain the deviation values. Determine whether each deviation value is within the design tolerance range. If so, the product to be tested is a good product. Mark each test parameter as a normal parameter and record the normal parameter values. If not, the product to be tested is a defective product. Extract the abnormal parameters from the test parameters and record them, then proceed to step S3.

3. The method for locating defective products in the electrolytic cell assembly process according to claim 2, characterized in that, In step S1, the construction process of the association rule base is as follows: S1.1, Construct the initial association rule base; S1.2 Collect historical data on defective products and construct a training dataset; S1.

3. Using the training dataset as input, the decision tree algorithm is used to train the initial association rule base and output the trained association rule base, thus completing the construction of the association rule base.

4. The method for locating defective products in the electrolytic cell assembly process according to claim 3, characterized in that: In step S5, if none of the candidate problem types in the association rule base are problems of defective products, then after manually identifying the new problem of the defective product, the new problem is recorded. When the new problem and the identified problems of defective products accumulate to a preset number, they are updated to the training dataset, and the association rule base is retrained to complete the update of the association rule base.

5. The method for locating defective products in the electrolytic cell assembly process according to claim 4, characterized in that: In step S2.2, the values ​​of various detection parameters of the product to be tested in the electrolytic cell assembly process are obtained by using docking sensors or manually.

6. The method for locating defective products in the electrolytic cell assembly process according to claim 5, characterized in that: In step S2.3, when the product to be inspected is defective, if there is a critical inspection parameter whose deviation value is exactly the critical value, then the abnormal parameter and the critical inspection parameter are extracted and recorded, and then step S3 is executed. During the execution process, the critical inspection parameter is used as a reference to determine the candidate problem type; the critical value is one of the values ​​at both ends of the design tolerance range.

Images